[jdk17u-dev] RFR: 8269404: Base64 Encoding optimization enhancements for x86 using AVX-512

[Scott Gibbons]

On Tue, 8 Nov 2022 13:29:20 GMT, Goetz Lindenmaier <goetz at openjdk.org> wrote:

>> This is a backport of JDK-8269404: Base64 Encoding optimization enhancements for x86 using AVX-512. Backport is clean.
>> 
>> We have had numerous customer requests for this functionality to be backported due to the performance improvement.
>> 
>> There will be three additional PRs for which this is a dependency. This PR is dependent on PR #860.
>> 
>> Risk: I view the risk of this backport to be minimal. This code has been in use for many months with no bugs reported.
>> 
>> Testing: x86_64 build, affected tests, tier1
>> 
>> Thanks,
>> --Scott
>
> What is wrong with the pre-submit tests?
> In addition, I will run the five changes through our testing.

@GoeLin This seems not to be the way in which I would see dependent PRs handled.  It appears that the bot added the dependent PRs commit to this commit, which should not happen.  If the previous commit was merged into master, and this commit was retargeted to master, the previous PR should not be re-applied.  It is already in master and should not appear here.

I may not fully understand the dependent PR process, so clarification would be appreciated.  I'm referencing an email from Erik Helin and Robin ?? entitled "New Skara feature: dependent pull requests" dated Mar 23, 2021.

As can be seen from this PR's history, the patch was clean before the previous PR was integrated.  Then once the previous was integrated, this PR shows conflicts.  I did this manually (checkout openjdk/jdk17u-dev:master, then backported the patch) which worked as advertised.

Please help.  Thanks.

-------------

PR: https://git.openjdk.org/jdk17u-dev/pull/861